Readers of the Web Analytics Report know that there are many steps to creating a successful web analytics initiative -- such as developing a strong organization, and creating processes to ensure that analytics get integrated into the mainstream of a company's operations.

If you're a web analytics manager for a large organization, one of the biggest challenges you likely face in creating enterprise-wide standards is at the very heart of web analytics: data collection.

Typically lots of energy goes into data accuracy, but it's easy to forget that it accuracy is tied to complete data collection. Most issues of data accuracy can be traced back to tags not being applied to all pages, or within applications, or that variables aren't set up to capture the required data.

As web analytics becomes more critical to senior managers, the need to create "apples to apples" comparison of site performance among different regions, different countries, or multitudes of company divisions increases in importance. To develop this level of macro-reporting, you need to make sure that the data collected from your Japanese site maps to the same variables as the data from your site in Spain, the USA, and so on. This will give you the confidence to develop roll-up views of activity across the entire enterprise, as well as individual site performance reports that may be judged through use of the same metrics and with a data set that all stakeholders agree to be complete and accurate.

To address the issue of complete data collection across multiple sites, I'm finding an increasing number of enterprises spending time developing a global tagging standard. Simply put, this tag includes the key common variables required for standardized data collection to provide agreed upon management metrics. Local teams may then customize the tag with additional variables if they require more advanced reporting. I'm seeing this especially with international implementations of Omniture. One of the gotchas a global tagging standard of SiteCatalyst helps avoid is the possibility of sites using different variable mappings, which could yield very different report results.

Think about it: standard data collection drives standard reporting, which you then can base standard report consumer training, which can then drive enterprise-wide agreement about web analytics definitions and results. Makes a lot of sense to me...what do you think?